POLITICS: Campaigns

Washington Post: So our election model is pretty accurate and it predicts the GOP takes the Senate --- but what the hell do we know?

By |
Beltway Confidential,Opinion,The Washington Post,2014 Elections,Campaigns,Becket Adams

The Washington Post covers the nation's most important elections with the help of Election Lab, a statistical model built by three political scientists.

And the model is pretty good, accurate even, according to politics writer Chris Cillizza.

This same election model currently predicts that the the Republican Party will retake the U.S. Senate after the November midterm elections. In fact, Election Lab puts the GOP's chances of picking up six Senate seats at around 86 percent.

But this, according to Cillizza, could be wrong.



The Washington Post's election model, which is supposedly pretty accurate, gives the GOP an 86 percent edge going into the midterms, so now it's up to Cillizza to explain the ever-so-fickle nature of … election models.

“Models are, by their nature, data-driven. (That's why models tend to get better the closer the election gets. There's just more raw material — poll numbers, fundraising numbers etc. — to mine.) Because of that reality,” he writes, “models tend to favor elements of races that can be easily quantified (presidential approval, GDP growth, fundraising) and diminish less easily quantifiable factors like candidate quality and the sort of campaigns being run on the ground.”

Tell me more about this black magic.

“How a candidate does on the stump, how they come across in TV ads, how smartly they are spending their money, how cohesive their campaign team is, how on message they can be — all of these things matter to the final outcome of races,” he writes.

Oh. So maybe data-driven models, like the one Election Lab has put together, are not that reliable.

“They don't decide the final outcome but neither are they irrelevant in it,” Cillizza continued, noting the oddity of Sen. Heidi Heitkamp, D-N.D., and her 2012 campaign victory.

Look, all Cillizza's trying to say is that data models are good — but they're not everything.

So even if Election Lab is pretty good, as we have been led to believe, and it predicts a strong November for Republicans, just remember: It could be wrong.

“Models need to be understood for what they can tell us — and what they can't,” he writes. “What the Election Lab model tells us is that the environment is very ripe for a Republican takeover of the Senate. What it doesn't is how the specifics of each campaign and candidate can change that dynamic in small but ultimately potential important ways.”

We're convinced! We'll ignore Election Lab from now on and seek out other models that don't leave us wracked with so much doubt. I mean, the Washington Post's model could be accurate, predicting large Republican gains, but it could also be wrong. Who can really know?

Thanks for the warning, guy!

You can read more about Election Lab's methodology here, which, again, Cillizza assures us is pretty solid.

View article comments Leave a comment

More from